data format for fasttext python

Question

I want to build and test classifier. I made 2 files - train.txt and test.txt

---

train.txt/test.txt

---

__label__1 , (text)\\n

---

__label__2 , (text)\\n

---

__label__3 , (text)\\n

---

and so on. After building a classifier:

classifier = fasttext.supervised("train.txt","model")

result = classifier.test("test.txt")

The precision is always 1/n (n - number of lines). So the classifier is for every input returning the same entity (for example 2 for all inputs).

Answer 1

You don't need to have a comma between a label and a text. All label should be prefixed (as in your example) and be before the text in both train and test files.